It doesn't look like any of the TTS solutions are sufficient. I'll be removing the language limitation sometime next year, once the interfaces are built out for custom audio and custom annotations.

If a good Dzongkha TTS comes out, then I'd be happy to integrate it natively!